ScienceDirect. Comparative fatigue resistance of overhead conductors made of aluminium and aluminium alloy: tests and analysis

Transcription

1 Available online at ScienceDirect Procedia Engineering 133 (2015 ) th Fatigue Design conference, Fatigue Design 2015 Comparative fatigue resistance of overhead conductors made of aluminium and aluminium alloy: tests and analysis R.B. Kalombo a, J.M.G. Martínez a, J.L.A. Ferreira a, C.R.M. da Silva a, J.A. Araújo a,* a University of Brasília, Department of Mechanical Engineering, Brasília, CEP Abstract This paper presents a comparative study for the fatigue behavior between an All Aluminium Alloy Conductor (AAAC) and a Aluminum Conductor Steel Reinforced (ACSR). These cables use Aluminium Alloys (AA) 6201-T81 and 1350-H19, respectively. This test program was carried out at a resonant fretting fatigue bench for overhead conductors at the University of Brasília. The results showed that the ACSR conductor fails by fatigue when is subjected to bending amplitude stress levels which are in average more than 1.25 times greater than that sustained by AAAC conductors (for the life ranges studied). In terms of durability, the ACSR has a fatigue life 5 times greater than the AAAC. A notch sensitivity analysis using Neuber s approach was invoked to try to explain this anomalous behavior for the cables fatigue performance, considering that the AA 6201-T81 has a significantly higher fatigue limit than the AA H Published The Authors. by Elsevier Published Ltd. This by Elsevier is an open Ltd. access article under the CC BY-NC-ND license ( Peer-review under responsibility of CETIM. Peer-review under responsibility of CETIM Keywords: Fatigue, fretting fatigue, ACSR Conductor, AAAC Conductor, transmission line. 1. Introduction The conductor is perhaps the most important component in transmission lines. It conducts the electricity and its contribution to the total cost of the line is significant. Therefore, the design and the maintenance of this component is vital for transmission line companies. One of the mechanical problems faced by power line conductors during their operation is fretting fatigue due to the Aeolian vibration. Fretting is caused by the small relative movement between wires and by the contact between the conductor and the suspension clamp (or other hardware such as spacer clamps or dampers, etc) [1,2]. Aeolian vibration is caused by the vortices produced due to fluid (air)-structure (conductor) interaction. The vortices formed alternately from the top and bottom of the conductor cause its bending in the hardware attachment points like in the suspension clamp [1,2]. This bending stress generated by the aeolian vibration, associated with the mean stretching load of the conductor, may provoke the fatigue of wires. Even though Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license ( Peer-review under responsibility of CETIM doi: /j.proeng

2 224 R.B. Kalombo et al. / Procedia Engineering 133 ( 2015 ) the fatigue phenomenon appears to be a more well-understood problem for some structures, it is clearly not the case for overhead conductors [3]. Traditionally, most high voltage transmission lines in the world have been constructed using the ACSR conductors stretched with Every Day Stress (EDS) between 18 and 20% of the conductor Rated Tensile Strength (RTS). The preference for the ACSR conductor is primarily because the steel wires in the internal conductor layer allow it to sustain higher tensile loads than pure aluminium conductors. The use of this EDS range is to ensure a long fatigue life due to a low static and dynamic mechanical stress. All Aluminium Alloy Conductors (AAAC) have been considered as an alternative to the ACSR as the absence of a steel core allows the AAAC conductor to transmit a larger amount of energy than the ACSR for the same conductor diameter. Additionally, the aluminium wires of the AAAC present better mechanical properties than those in the ACSR. Therefore, the AAAC conductors have been strung at higher tensions and longer spans than traditional ACSR. However, there are few data and publication on the fatigue behavior of AAAC. This paper will present a comparative study on the fatigue behavior between AAAC and ACSR conductors. 2. Fundamental theories 2.1. Fatigue of power line conductor and its S-N graph The fundamental cause of conductor fatigue failure is the cyclic bending stress due to aeolian vibration. This bending stress has been estimated between 200 and 1400 micro strains [3]. This type of failure occurs in the regions where there are restrictions on the vibratory motion of the conductor, such as the suspension clamps, spacers and dampers [1-6]. This restriction causes the conductor s wires to slip across each other and the friction forces from this relative movement cause fretting between the wires and the contact area between the conductor and restriction component like the suspension clamp. Once crack-induced fretting has formed, it may lead to fatigue and eventually, to a brutal facture of the conductor [7]. One way to characterize the material in engineering related to the fatigue life is the application of Wöhler s graph (S-N graph). This graph correlates the number of cycles required to break a specimen subjected to a nominal cyclic stress. Due to its complex configuration the wire stress generated by bending of the pre-loaded conductor and used to construct the Wöhler s graph is obtained by the Poffemberger and Swart (P-S) formula [8]. This mathematical expression has been adopted by IEEE as a standard for the measurement of vibration in conductors since Another feature of the Wöhler s graph for conductors is associated with the criterion selected to characterize the conductor failure. One criterion established by CIGRÉ [4] considers that the conductor fails by fatigue when the number of wires broken is equal to either three or 10% of the total number of aluminium wires, whichever is greater. Because of the difficulty and the high cost involved in the conductor fatigue test, the CIGRE committee proposed a safety limit line, well-known as the CIGRÉ Safe Boder Line (CSBL). This permits the companies involved in designing transmission lines to determine the conductor s life without making a substantial investment. The CSBL is a S-N curve derived from the diverse fatigue tests made on various components such as conductors and aluminium wires [2]. This graph is considered to be a very conservative limit on conductor fatigue life. The CSBL equation can be expressed as follows: (1) Where σ a is the stress amplitude in MPa; N i is the fatigue life in number of cycles for a specific stress; A and B are material constants. Table 1 reports the values of A and B as a function of the fatigue life and of the number of layers of the conductor Poffemberger-Swart formula Based on an analogy with Euler beam theory (Fig. 1), Poffemberger and Swart [8] developed a formula which establishes a relation between the displacement and the bending stress amplitude for an aluminium wire in the external layer of the conductor:

3 R.B. Kalombo et al. / Procedia Engineering 133 ( 2015 ) (2) Table 1. Constants (A and B) of the CIGRÉ Safe Border Line (CSBL) Number of cycles N O of layers of aluminium wire conductor N < 2x10 7 N > 2x10 7 A B A B > Figure 1. Schematic figure of the conductor/clamp assembly showing the standard position to measure the bending displacment Yb and the local on the aluminium wire of the external layer where the stress amplitude is computed. Where σ a is the dynamic bending stress amplitude (zero to peak), Y b is the bending displacement range (peak to peak) of the conductor measured vertically at a position on the cable 89 mm from the last point of contact between the conductor and the cable (LPC) and K is the constant of Poffenberger: (3) The parameters E a [MPa], d [mm], x and p are, respectively, the Young's modulus for aluminium, the diameter of the aluminium wire of the outer layer of the conductor, the distance on the conductor between the last point of contact (LPC) and the position where the vertical displacement (usually x = 89 mm) is measured and p is given by: (4) being T [N] the static conductor tension (Every Day Stress: EDS); and EI [N mm 2 ] is the conductor's flexure stiffness whose expression is given in Eq. 5: (5) Where n a, E a, and d a are, respectively, the number, the Young's modulus and the diameter of an aluminium wire; and n s, E s, and d s are, respectively, the number, the Young's modulus and the diameter of a steel wire Characteristics of the aluminium alloys used in the fabrication of ACSR and AAAC As presented above, the aluminium alloys 1350-H19 and 6201-T81 are used, respectively, for the production of the ACSR and the AAAC conductors.

4 226 R.B. Kalombo et al. / Procedia Engineering 133 ( 2015 ) AA 1350-H19 The AA 1350 is a series 1xxx alloy. This series has not experienced a heat treatment and have a strict control on the percentage of impurities, as they can dramatically reduce the electrical conductivity of the material. As a result of this control, the electrical conductivity is equal to 62% (International Annealed Copper Standard -IACS). This essentially means that the wire has an electrical conductivity equal to 62% of the conductivity of a wire copper with the same diameter. On top of that, the AA 1350 presents a high corrosion resistance, good formability, good weldability and good response to anodizing. Naturally, this alloy is ductile but there is an extra hardening due to the cold hardening process (H19) suffered by the material to produce the thread. The chemical composition, as well as the mechanical properties of this alloy are reported in Tables 2 and AA 6201-T81 The AA 6201 is a 6xxx alloy. This series contains magnesium and silicon as their main alloying elements. The combination of these alloying elements results in the intermetallic element Mg 2 Si, which is responsible for hardening before reaching equilibrium in the over-aging. When the treatments of this alloy are made properly, the presence of Mg 2 Si favors the formation of fine precipitates homogeneously distributed in the material, which results in a substantial increase in the hardness. The electrical conductivity of the AA 6201-T81 is equal to 52%, and though not as good as the one for AA It has a higher mechanical strength and resistance to corrosion than the AA The AAAC conductors are hence recommended for lines close to the coast and/or requiring a higher mechanical strength than obtained with the ACSR conductor. The AA 6201 undergoes a solubilization treatment, cold working and artificial aging to achieve a specific degree of stretch (Classification T81). Tables 2 and 3 report the chemical composition and the mechanical properties of the AA 6201-T81, respectively. Table 2. Chemical composition of AA 1350-H19 and AA 6201-T81 [9,10] All Alloy Al B Cr Cu Ga Fe Mn Mg Si V+Ti Zn Cada Total AA 1350-H AA 6201-T Table 3: Basic mechanical properties of AA 1350-H19 and AA 6201-T81 [9,10] Limit of Resistance All Alloy Young's Modulus [GPa] Yield Strength σ rt [MPa] Ultimate Strength σ y [MPa] Fatigue σ e [MPa] Hardness [HBn] Strain (Deformation) at rupture [%] AA 1350-H AA 6201-T Fatigue strength reduction factor and notch sensitivity In fatigue analysis, it is important to define the effect of geometrical discontinuities on the fatigue strength of the material. This can be done by introducing the concept of a strength reduction factor and of notch sensitivity. The fatigue strength knock-down factor K f, is an empirical parameter which attempts to evaluate the effect of the stress concentration on the fatigue strength of the piece. K f is defined as: (6) where σ f is the fatigue limit of the plain specimen and σ notch is the corresponding limit for a notched specimen. The notch sensitivity factor, q, is another empirical parameter which quantifies the effect of the stress concentrator on the material fatigue resistance. In general, when the material is more ductile, its sensitivity to notches is lower

5 R.B. Kalombo et al. / Procedia Engineering 133 ( 2015 ) and brittle materials are more sensitive to discontinuities. The ductility and the brittleness of metals are strongly correlated to their strength and hardness. Therefore, materials with a low strength and less hardness have a tendency to be less sensitive to discontinuities, while the ones with high strength and hardness have a tendency to be more sensitive. Notch sensitivity depends not only on the mechanical characteristics of the material but also on the notch radius. It was found empirically that as the notch radius tends to zero, the sensitivity of the material decreases and also approaches zero. Several classic models have been developed to determine the sensitivity of the material to the presence of notches. However, the model developed by Neuber [11] which is written in terms of Neuber s constant ( ) and the notch radius (r) seems to be the most commonly used. (7) The Neuber's constant is a parameter that depends on the material type and it is related to the tensile strength of the material. Figure 2 shows the behaviour of the Neuber's constant as a function of the tensile strength for two aluminium alloys, annealed and hardened. The aluminium alloys in this study are hardened alloys, hence Neuber's constant for AA 6201-T81 and AA 1350-H19 can be obtained from this graph (Fig. 2) and are and 1.517, respectively [12]. It is possible to evaluate the notch sensitivity for these two materials by replacing these values in Eq. 7 and by writing the expression as a function of the square root of the notch radius. a rt a rt Figure 2. Behaviour of the Neuber s constant as a function of the tensile strength of the aluminium alloys (Adapted from [12]) Correction of the fatigue resistance as function of the fatigue knock-down factor From Eq. 6, one can obtain an expression which allows us to evaluate the fatigue resistance of the notched component, σ notch. This requires the fatigue resistance of material, σ f, and the fatigue knock-down factor, K f, which depends on the stress concentration factor, : (8) being defined as: (9) where σ max is the local stress observed in the hot spot and σ nom is the nominal stress.

6 228 R.B. Kalombo et al. / Procedia Engineering 133 ( 2015 ) Substituting (6) in (8), σ notch becomes (10) 3. Experimental programme 3.1. Materials For this research, one sample of an ACSR conductor (named A) and two samples of AAAC conductors (named B and C) were considered. All tests were performed using mono-articulated suspension clamps suitable to each conductor diameter. The conductor Rate Tensile Strength (RTS) and the Poffenberger-Swart s constant for the conductors tested are reported in Table 4. The respective stretching loads (Every Day Stress: EDS) used for the test are also presented in Table 4. For all tests, the EDS have been limited to 20% of the RTS. Table 4. The Rate Tensile Strength (RTS) and geometrical properties of the conductors Conductor Identification Conductor Type Rate Tensile Strength (RTS) [kgf] Material, Number and Diameter of Steel wire [mm] Aluminium Constant of Poffenberger [MPa/mm] Stretching loads EDS[kgf] A ACSR x x B AAAC x C AAAC x The fixation of the conductor on the bench has been made by using a mono-articulated suspension clamp with the diameter ranging between 17 to 27 mm for samples A and B and between 28 to 38 mm for sample C Apparatus used for the experimental work A series of tests have been conducted in the resonant bench of the Laboratory of Fatigue and Structural Integrity of Overhead Conductors in the University of Brasília (UnB). The schematic drawing of the test bench is shown in Fig. 3. The active span for the test bench was set to 40 m length. The laboratory temperature is controlled to avoid changes in the conductor s tension load as a result of temperature variations. More details on the description of the laboratory as well as of the bench can be found elsewhere [7,11] Experimental Procedure Figure 3. Scheme of Conductor Fatigue Test Bench at the University of Brasilia (UnB). The fatigue performance of the three conductors has been compared based on the bending stress (and not on ).

7 R.B. Kalombo et al. / Procedia Engineering 133 ( 2015 ) During the test, the bending amplitude was measured on the conductor by means of a accelerometer located at 89 mm from the LPC between the conductor and the suspension clamp. The bending (nominal) stress,, in the aluminium wire of the conductor s outer layer in a point diametrically opposite to the LPC has been found by using the Poffenberger-Swart s equation (Eq. 2). One should bear in mind that this bending stress amplitude is a function of the conductor s displacement amplitude at a point 89 mm from the LPC, which is test control variable. This means that such a bending stress has also been controlled during the entire fatigue test. The different stress amplitudes considered in the experimental program and used to plot the S-N graph are reported in Table 5, as well as their respective bending amplitude at 89 mm from the LPC. Table 5: Bending stress at Last Point of Contact (LPC) and bending amplitude at 89 mm from LPC Condutor Bending Stress, P-S [MPa] , A 0.9 1, Bending Amplitude at 89 mm B from LPC [mm] C , Results and Discussions Figure 4 shows the S-N curves generated for these fatigue tests. Again the termination criterion was the one suggested by CIGRE, i.e., tests were interrupted after 3 aluminium wire breaks or after 10% of the number of Aluminium wires were broken, whichever is greater. The qualitative analysis of these results indicates that the AAAC conductors (B and C) have significantly less fatigue strength than the ACSR conductor (A). It can also be observed that the results of these experimental points obtained for the conductors B and C are apparently very close one to the other. This may indicate that these data sets could represent the same fatigue behaviour for the conductors B and C. In order to verify this hypothesis, the life data sample for conductors B and C were classified by stress levels and the statistic t-student law was applied. It has statistically been noted that usually fatigue life follows the log-normal distribution. The statistical analysis shows that data from the fatigue life of conductors B and C could be considered as data from the same sample. Figure 4. S-N graph for the conductor A, B and C. The two AAAC conductors (B and C) used in this study have different geometrical characteristics the number of

8 230 R.B. Kalombo et al. / Procedia Engineering 133 ( 2015 ) layers and aluminium wires, the overall diameter and therefore the suspension clamps used for testing these conductors (B and C) have different dimensions. In addition to these, all conductors tested (A, B and C) have been stretched to the same loads (20% EDS) for the duration of the experiment. Therefore, one can could in principle infer that the smaller resistance to fatigue of AAAC is primarily caused by the stress concentration sensitivity of the AA 6201-T81. Assuming that the experimental fatigue data obtained for conductors B and C represent the typical behaviour of AAAC conductors, then their life data could be expressed by a single S-N curve (Fig. 5). In order to carry out a comparative analysis in terms of fatigue performance not only between the conductors tested, but also with the CSBL, the fatigue strength for the data set produced for the Cables A, B and C have been normalised with respect to the CSBL stress (Fig. 6). In Figure 7 these data was regrouped considering again that cables B and C have a fatigue behavior typical of AAAC conductors and then plotted together with the ACSR data. Further, a line of the fatigue resistance ratio between the ACSR and AAAC conductors for different lives was added. It seems clear from Figs. 6 and 7 that, for a life ranging from 10 6 to 10 7 cycles, the fatigue strength of the ACSR is 50 to 60% greater than those estimated by the CSBL, while the fatigue strength for the AAAC are only around 20 to 40% greater. It can also be seen that for smaller lives (10 6 cycles) the fatigue resistance of the ACSR is approximately 40% greater than the one observed for AAAC. This additional fatigue resistance of the ACSR with respect to the AAAC nevertheless tends to reduce for larger lives. Figure 5. S-N graph of different conductor s type (ASCR and AAAC). Figure 6. Normalized S-N graphs for the conductor A, B and C in relation of the CSBL stress.

9 R.B. Kalombo et al. / Procedia Engineering 133 ( 2015 ) Discussion and Conclusions A comparison between the mechanical properties of the AA 6201-T81 and the AA 1350-H19 (Table 3) revealed that the AA 6201-T81 has a higher yield and ultimate strength than the AA 1350-H19. Consequently, its fatigue resistance also proved higher than the one for the AA 1350-H19, as it is usually the case. Therefore, one could in principle expect that the AAAC conductor (made of AA 6201-T81) would also have a higher fatigue strength than the ACSR (AA 1350-H19). However, care must be exercised as, due to the contact loads and the fretting between wires and between the wires of the outer layer and the suspension clamp, a complex stress state (with stress concentration) and micro notches may arise. Indeed, as just shown in the previous section, the fatigue strength of the AAAC conductor was significantly lower than that observed for the ACSR conductor. In this setting, notch sensitivity may well explain this behavior. Figure 7. Normalized S-N graphs for the ACSR and AAAC conductors in relation of the CSBL stress and their fatigue resistance ratio for different lives. From Eq. 10, one can observe that the fatigue resistance of a conductor is a function not only of the fatigue resistance of the wire material, but also of the notch sensitivity and the stress concentration factor associated to discontinuities, grooves (notches) and fretting marks. Figure 8(a) shows the variation of notch sensitivity with notch radius, r, for the two aluminium alloys present in AAAC and ACSR cables. The analysis of these graphs allows to verify that, for the same r, the AA 6201-T81 has a greater notch sensitivity than the AA 1350-H19. An alternative way to visualize this behavior is shown in Fig. 8(b). Such a graph depicts the ratio between the notch sensitivity for AA 6201-T81 and for 1350-H19 against r. It is clear from this graph that for small values of the notch radius the notch sensitivity for the AA 6201-T81 ( ) is significantly higher than the one for AA 1350-H19. For instance, for this ratio is around This article discussed the theoretical aspects associated with the effect of notch sensitivity on the fatigue strength of the aluminium alloys AA 6201-T81 and AA 1350-H19 used respectively for the manufacturing of AAAC and of ACSR conductors. In addition, a comparative test program was conducted to determine the life of these two conductor types: AAAC and ACSR. The following main conclusions can be drawn from this work: The fatigue performance of the AAAC conductor is qualitatively much lower than that observed for the ACSR conductor; For a same life, the ACSR conductor failed when subjected to bending stresses which were 1.25 times greater than the ones necessary to break the AAAC conductor; For the same stress amplitude, the ACSR conductor has, in average, a fatigue life five times greater than the one for the AAAC; and The greater notch sensitivity of AA 6201-T81 with respect to AA 1350-H19 is one of the possible reasons to explain such anomalous behavior.

10 232 R.B. Kalombo et al. / Procedia Engineering 133 ( 2015 ) q 6201-T81 /q 1350-H (a) Notch Radius, r [mm] (b) Figure 8. (a) The notch sensitivity and (b) the notch sensitivity ratio versus the notch radius for the two aluminium alloys. Acknowledgments The authors would like to acknowledge the financial support of TAESA, ATE II, ATE III, EATE, TME, AETE and Brasnorte for this research (contract number ) by means of the R&D program of ANEEL. The support of Finatec is also acknowledged. References [1] CIGRÉ-SCB2-WG11-TF , Fatigue endurance capability of conductor/clamp systems Update of Present Knowledge, [2] IEEE, Guide for Aeolian Vibration Field Measurements of Overhead Conductors, [3] J. Chan, EPRI Transmission Line Reference Book: Wind-Induced conductor motion, Electric Power Research Institute, Palo Alto, California, [4] CIGRE WG 04 SC 22-02, Recommendations for the evaluation of the lifetime of transmission line conductors, ELECTRA 63, March [5] CIGRE SC 22, Final Report of Working Group 04, Endurance Capability of Conductors Inside Clamps, Paris, Jul 1988, 19p. [6] C.B. Rawlings, Fatigue of Overhead Conductors, Transmission Line Reference Book, Electrical Power Research Institute, Palo Alto CA, 1979, pp [7] A.A. Fadel, D. Rosa, L.B. Murça, J.L.A. Ferreira, and J.A. Araújo, Effect of high mean tensile stress on the fretting fatigue life of an Ibis steel reinforced aluminium conductor, International Journal of Fatigue, Volume 42, September 2012, pp [8] J.C. Poffenberger, R.L. Swart, Differential Displacement and Dynamic Conductor Strain, IEEE Transactions Paper, Vol. PAS 84, 1965, pp [9] Norma ASTM B398/B398M 02 (Reapproved 2007), Aluminum-Alloy 6201-T81 Wire for Electrical Purposes, Copyright ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA , United States. [10] B.L. Silva, J.A. Araújo, J.L.A. Ferreira, High-cycle notch sensitivity of alloy steel ASTM A743 CA6NM used in hydrogenator turbine components, Frattura ed Integrità Strutturale, v. 14 (2010) pp [11] V.F. Volker, R.B. Kalombo, R.M.S. Cosme, M.N. Nogueirab, J.A. Araújo, Effect of chromium nitride coatings and cryogenic treatments on wear and fretting fatigue resistance of aluminium, v. 116 (2014), pp [12] P. Kuhn, I.E. Figge, Unifed notch-strength analysis for wrought alominium alloys, NASA TN D-125 9, pp. 54.